CN109978859B - Image display adaptation quality evaluation method based on visible distortion pooling - Google Patents

Image display adaptation quality evaluation method based on visible distortion pooling Download PDF

Info

Publication number
CN109978859B
CN109978859B CN201910237280.0A CN201910237280A CN109978859B CN 109978859 B CN109978859 B CN 109978859B CN 201910237280 A CN201910237280 A CN 201910237280A CN 109978859 B CN109978859 B CN 109978859B
Authority
CN
China
Prior art keywords
image
display
local
pooling
original image
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201910237280.0A
Other languages
Chinese (zh)
Other versions
CN109978859A (en
Inventor
牛玉贞
陈钧荣
张帅
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Fuzhou University
Original Assignee
Fuzhou University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Fuzhou University filed Critical Fuzhou University
Priority to CN201910237280.0A priority Critical patent/CN109978859B/en
Publication of CN109978859A publication Critical patent/CN109978859A/en
Application granted granted Critical
Publication of CN109978859B publication Critical patent/CN109978859B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10004—Still image; Photographic image
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10016—Video; Image sequence
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20021—Dividing image into blocks, subimages or windows
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30168—Image quality inspection

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Quality & Reliability (AREA)
  • Image Analysis (AREA)
  • Testing, Inspecting, Measuring Of Stereoscopic Televisions And Televisions (AREA)

Abstract

The invention relates to an image display adaptation quality evaluation method based on visible distortion pooling. The method comprises the following steps: s1: reading an original image, and generating a saliency map S by using a salient object detection algorithm; s2: establishing a pixel level mapping relation between an original image and a display adaptive image by using an image registration method; s3: calculating a block-level fidelity F of the image using the pixel-level mapping relationship; s4: and using the visible distortion pooling strategy pooling saliency map S and image fidelity F to calculate and display the overall objective evaluation quality Q of the adaptive image. The invention can effectively simulate and display the human eye evaluation mechanism in the adaptive image objective quality evaluation, solves the defect of the importance degree weighting pooling strategy in comparison among groups, is beneficial to improving the consistency between the evaluation score and the subjective score of the objective evaluation method, and can be applied to the field of objective quality evaluation for displaying adaptive images and other application fields needing to adopt the pooling strategy to combine with local information to evaluate the overall quality.

Description

Image display adaptation quality evaluation method based on visible distortion pooling
Technical Field
The invention relates to the field of image and video processing and computer vision, in particular to an image display adaptation quality evaluation method based on visible distortion pooling.
Background
In recent years, researchers have proposed many image display adaptation methods to operate on the content of an image to adapt to various target display screens having different sizes and aspect ratios. However, at present, there is no image display adaptation method that can be well adapted to various application scenarios and generate satisfactory display adaptation images. There are many factors that restrict the further development of the image display adaptation method, and one important factor is that there is still no image display adaptation quality assessment (IRQA) method that is consistent with human subjective perception. Most image display adaptation methods tend to employ small scale subjective tests on the data set to verify the effectiveness of the method. Although the results of subjective testing can intuitively illustrate how good a displayed adapted image is in visual quality, subjective testing methods are often non-automated, have specific requirements on the quality of the tester, and are not efficient in dealing with large-scale data. Therefore, it is important to develop an objective and effective IRQA method to promote further development of image display adaptation methods.
Early IRQA metrics such as bilateral similarity (BDS), color placement (CL), and land mobile distance (EMD) evaluated the objective quality of display-adapted images by simply calculating image distance or converting the display-adapted image and a reference image into feature descriptors with the same dimensions. However, these methods do not take into account differences in sensitivity of human subjective evaluation to content deformation of different images, and thus objective evaluation results have low consistency with human subjective perception.
In recent years, researchers at home and abroad propose a plurality of solutions to the difficulties of IRQA. Liu et al uses global structural similarity and local correspondences to evaluate the display adaptation image. They first extract global geometry through a traversal from coarse scale to fine scale and build local pixel correspondences in a top-down manner. Secondly, an importance map is generated by combining low-level significance information and high-level face information. Finally, based on the correspondence, a visual quality of the display adapted image is evaluated using an importance weighted similarity measure. Fang et al established dense correspondences between images using SIFT-Flow and computed Structural Similarity (SSIM) maps of different scales to measure and display structural information retained in the adapted images. They generate visual importance maps in conjunction with bottom-up and top-down image saliency information, and then evaluate the overall visual quality of the display-adapted images by computing importance-weighted SSIM maps. Zhang et al proposed a SIFT-Flow based backward registration algorithm to model the geometric changes experienced by the original image during display adaptation. The visual quality of the display adapted image is then evaluated by calculating an importance weighted Aspect Ratio Similarity (ARS) metric. These methods all require the integration of a visual attention model, and the pooling method is used to pool the overall visual quality according to the local image quality. However, the similarity measure for measuring local quality is still insufficient, and the common Importance Weighted Pooling (IWP) strategy is also prone to the problem of changing the importance rule in comparison among groups, so the consistency between the objective evaluation result and the subjective score is still not high.
Disclosure of Invention
In view of this, the present invention provides an image display adaptive quality assessment method based on pooling of visible distortions, which is helpful for improving the consistency between objective quality assessment results and subjective perception, and has an obvious optimization effect.
The invention is realized by adopting the following scheme: an image display adaptation quality assessment method based on visible distortion pooling comprises the following steps:
step S1: acquiring and reading an original image, and generating a saliency map S by using a salient object detection algorithm;
step S2: acquiring a display adaptive image, and establishing a pixel-level mapping relation between the original image and the display adaptive image by using an image registration method based on SIFT-Flow;
step S3: averagely dividing the original image into a plurality of local blocks; averagely dividing the display adaptive image into a plurality of local blocks corresponding to an original image;
step S4: calculating the block-level fidelity F of the display adapted image by using the pixel-level mapping relation obtained in the step S2 to obtain the image fidelity F of the display adapted image relative to the original image;
step S5: pooling said saliency map S and said image fidelity F using a visible distortion pooling strategy and calculating an overall objective assessment quality Q of said display adapted imageNBP。
Further, the step S2 specifically includes the following steps:
establishing a pixel-level mapping relationship between the original image and the display-adapted image by using an SIFT-Flow-based image registration method, wherein an SIFT-Flow vector field from the display-adapted image to the original image can be calculated by solving the following energy minimization problem:
Figure BDA0002008848380000031
wherein, IoAnd IrRespectively representing an original image and a display adapted image; p is a radical ofrRepresenting a pixel coordinate in the display adapted image; q. q ofrIs prOne adjacent pixel coordinate of (a); epsilon represents a four-connected neighborhood set; w (p)r)＝(u(pr),v(pr) ) represents prSIFT-Flow displacement vector of (2); u and v represent the horizontal and vertical components of the SIFT-Flow displacement vector, respectively; d is a threshold value, with a default value of 40; α is a weighting factor for the second term and has a default value of 2.
Further, the step S4 specifically includes the following steps:
step S41: calculating the aspect ratio similarity of the corresponding local blocks between the original image and the display adaptive image by using the pixel-level mapping relation, thereby calculating the geometric distortion of the local blocks, wherein the calculation formula is as follows:
Figure BDA0002008848380000041
wherein k represents the index sequence number of the local block in the original image; the ratio of variation of the width and height of the local block is rw(k)＝w*(k) N and rh(k)＝h*(k)/N；w*(k) And h*(k) Width and height, respectively, of local blocks of the reconstructed display-adapted image, obtained by calculating the maximum horizontal and vertical distances between pixels in the local blocks; when the variation ratio of width and height is the same, the local block equivalent to the original image is scaled in equal proportion, and the AR measure reaches the maximum value, i.e. SAR＝1；
Step S42: the area similarity of the corresponding local blocks is calculated by utilizing the pixel-level mapping relation, so that the information loss of the local blocks is measured, and the calculation formula is as follows:
Figure BDA0002008848380000042
wherein r isa(k)＝a*(k)/N2Is the area change ratio of the kth local block; η is a positive number, with a default value of 0.3, used to balance the geometric distortion and the weight of information loss in the image fidelity measure;
step S43: and calculating the image fidelity of the local block displaying the adaptive image by using the aspect ratio similarity and the area similarity, wherein the calculation formula is as follows:
F(k)＝SAR(k)·SA(k)。
further, the step S5 specifically includes the following steps:
step S51: defining distortion visibility of each local block according to the average value of each local block in the saliency map S, and calculating the objective quality score of each block according to the fidelity and the distortion visibility of the block:
Figure BDA0002008848380000051
wherein the content of the first and second substances,
Figure BDA0002008848380000052
an average saliency value representing the kth local block of the saliency map S; if image fidelity F (k) is normalized to [0,1 ]]Then the mass range of the kth local block is
Figure BDA0002008848380000053
Therefore, when
Figure BDA0002008848380000054
The mass range of the corresponding local block is [0,1 ]](ii) a And following with
Figure BDA0002008848380000055
The quality base of each local block is continuously changedLarge, the range is continuously reduced;
step S52: finally, obtaining an objective quality score of the whole display adaptive image by calculating the average value of the quality scores of all the local blocks; calculating the objective quality score of each local block according to the fidelity and distortion visibility of the block:
Figure BDA0002008848380000056
wherein Q isNBPRepresenting the objective quality score calculated by the visible distortion pooling strategy, and M representing the number of local blocks in the original image.
Further, in step S51, the defining the distortion visibility of each local block according to the average value of each local block in the saliency map S specifically includes: defining a mass range with small cardinality and large interval for local blocks with high significance; meanwhile, a quality range with large base number and small interval is defined for the local blocks with low significance.
Compared with the prior art, the invention has the following beneficial effects:
(1) the method is suitable for objective quality evaluation of display adaptive images, is beneficial to improving the consistency of objective quality evaluation results and subjective perception, and has obvious optimization effect.
(2) The method can be well suitable for display adaptive image quality evaluation, enables objective evaluation results and human subjective scores to keep better consistency, and can be used in the fields of display adaptive image quality evaluation, image display adaptive method optimization and the like.
Drawings
FIG. 1 is a flow chart of an embodiment of the present invention.
Fig. 2 is a flowchart of an implementation of the overall method according to the embodiment of the present invention.
Detailed Description
The invention is further explained by the following embodiments in conjunction with the drawings.
As shown in fig. 1 and 2, the present embodiment provides an image display adaptation quality assessment method based on visible distortion pooling, which includes the following steps:
step S1: acquiring and reading an original image, and generating a saliency map S by using a salient object detection algorithm;
step S2: acquiring a display adaptive image, and establishing a pixel-level mapping relation between the original image and the display adaptive image by using an image registration method based on SIFT-Flow;
step S3: averagely dividing the original image into a plurality of local blocks; averagely dividing the display adaptive image into a plurality of local blocks corresponding to an original image; (in this embodiment, the display adaptive image is completely divided into partial blocks of the same size, for example, 9 × 9 image, and the average is divided into 9 3 × 3 image blocks of the same size)
Step S4: calculating the aspect ratio and the area similarity of the corresponding local block by using the pixel-level mapping relation obtained in the step S2 to obtain block-level fidelity F;
step S5: pooling said saliency map S and said image fidelity F using a visible distortion pooling strategy and calculating an overall objective assessment quality Q of a display adapted imageNBP。
In this embodiment, the step S2 specifically includes the following steps:
establishing a pixel-level mapping relationship between the original image and a display adaptive image by using an SIFT-Flow-based image registration method, wherein an SIFT-Flow vector field from the display adaptive image to the original image can be calculated by solving the following energy minimization problem:
Figure BDA0002008848380000071
wherein, IoAnd IrRespectively representing an original image and a display adapted image; p is a radical ofrRepresenting a pixel coordinate in the display adapted image; q. q.srIs prOne adjacent pixel coordinate of (a); epsilon represents a four-connected neighborhood set; w (p)r)＝(u(pr),v(pr) ) represents prSIFT-Flow displacement vector of (1); u and v represent SI, respectivelyHorizontal and vertical components of the FT-Flow displacement vector; d is a threshold value, with a default value of 40; α is the weighting factor for the second term and has a default value of 2.
In this embodiment, the step S4 specifically includes the following steps:
step S41: calculating the aspect ratio similarity of corresponding local blocks between the original image and the display adaptive image by using the pixel-level mapping relation, thereby calculating the geometric distortion of the local blocks, wherein the calculation formula is as follows:
Figure BDA0002008848380000081
wherein k represents the index sequence number of the local block in the original image; the ratio of variation of the width and height of the local block is rw(k)＝w*(k) N and rh(k)＝h*(k)/N；w*(k) And h*(k) Width and height, respectively, of local blocks of the reconstructed display-adapted image, obtained by calculating the maximum horizontal and vertical distances between pixels in the local blocks; when the variation ratio of width and height is the same, the local block equivalent to the original image is scaled in equal proportion, and the AR measure reaches the maximum value, i.e. SAR＝1；
Step S42: the area similarity of the corresponding local blocks is calculated by utilizing the pixel-level mapping relation, so that the information loss of the local blocks is measured, and the calculation formula is as follows:
Figure BDA0002008848380000082
wherein r isa(k)＝a*(k)/N2Is the area change ratio of the kth local block; η is a positive number, with a default value of 0.3, used to balance the geometric distortion and the weight of information loss in the image fidelity measure;
step S43: and calculating the image fidelity of the local block displaying the adaptive image by using the aspect ratio similarity and the area similarity, wherein the calculation formula is as follows:
F(k)＝SAR(k)·SA(k)。
in this embodiment, the step S5 specifically includes the following steps:
step S51: defining the local blocks according to the mean value of each local block in the significance map S
Distortion visibility of partial blocks, the blocks being calculated from fidelity and distortion visibility of each block
Objective mass fraction:
Figure BDA0002008848380000091
wherein the content of the first and second substances,
Figure BDA0002008848380000092
an average saliency value representing the kth local block of the saliency map S; if image fidelity F (k) is normalized to [0,1 ]]Then the mass range of the kth local block is
Figure BDA0002008848380000093
Therefore, when
Figure BDA0002008848380000094
The mass range of the corresponding local block is [0,1 ]](ii) a And following with
Figure BDA0002008848380000095
The quality base number of each local block is continuously increased, and the range is continuously reduced;
step S52: finally, obtaining an objective quality score of the whole display adaptive image by calculating the average value of the quality scores of all the local blocks; calculating the objective quality score of each block according to the fidelity and distortion visibility of the block:
Figure BDA0002008848380000096
wherein Q isNBPRepresenting the objective quality score calculated by the visible distortion pooling strategy, M representing the original graphNumber of local blocks in the image.
In this embodiment, the defining the distortion visibility of each local block according to the average value of each local block in the saliency map S in step S51 specifically includes: a local block with high significance is defined with a small base number and a large quality range of an interval, so that the influence of significant image content on the overall quality of a display adaptive image is emphasized. Meanwhile, a mass range with large base number and small interval is defined for local blocks with low significance, so that insensitivity of a human visual system to distortion of insignificant contents is simulated.
In particular, the display-adapted image in the present embodiment is a distorted image obtained by modifying the image size, such as stretching and compression deformation, distortion similar to common blur, noise, and the like, and also belongs to a type of distorted image. In the present embodiment, the display-adapted image is acquired from the MIT ratebedme and CUHK data sets.
Preferably, the embodiment introduces an aspect ratio and an area similarity measure local quality in the display adaptive image quality evaluation, and pools the local quality and the saliency map by using a visible distortion pooling strategy to obtain the overall objective quality of the image. The local quality of the image is measured by adopting a method of combining the aspect ratio similarity and the area similarity, a new visible distortion pooling method is provided, and the human eye evaluation mechanism in the image quality evaluation can be better simulated and displayed while the local distortion is fully calculated. The method is suitable for a display adaptive image quality evaluation method, is beneficial to improving the consistency of objective quality evaluation results and subjective perception, and has an obvious optimization effect.
In particular, the present embodiment adequately measures the degree of distortion of the local image by calculating the aspect ratio and the area similarity. Second, a visible distortion pooling strategy is designed to pool the local image fidelity and saliency maps to compute the overall objective quality of the display-adapted image. The method can be well suitable for display adaptive image quality evaluation, enables objective evaluation results and human subjective scores to keep better consistency, and can be used in the fields of display adaptive image quality evaluation, image display adaptive method optimization and the like.
The above description is only a preferred embodiment of the present invention, and all equivalent changes and modifications made in accordance with the claims of the present invention should be covered by the present invention.

Claims (2)

1. An image display adaptation quality assessment method based on visual distortion pooling is characterized by comprising the following steps: the method comprises the following steps:
step S1: acquiring and reading an original image, and generating a saliency map S by using a salient object detection algorithm;
step S2: acquiring a display adaptive image, and establishing a pixel-level mapping relation between the original image and the display adaptive image by using an image registration method based on SIFT-Flow;
step S3: averagely dividing the original image into a plurality of local blocks; averagely dividing the display adaptive image into a plurality of local blocks corresponding to an original image;
step S4: calculating the block-level fidelity F of the display adapted image by using the pixel-level mapping relation obtained in the step S2 to obtain the image fidelity F of the display adapted image relative to the original image;
step S5: pooling said saliency map S and said image fidelity F using a visible distortion pooling strategy and calculating an overall objective assessment quality Q of said display adapted imageNBP；
The step S2 specifically includes the following steps:
establishing a pixel-level mapping relationship between the original image and the display-adapted image by using an SIFT-Flow-based image registration method, wherein an SIFT-Flow vector field from the display-adapted image to the original image can be calculated by solving the following energy minimization problem:
Figure FDA0003623317920000011
wherein, IoAnd IrRespectively representing an original image and a display adapted image; p is a radical of formularRepresenting an image in a display adapted imageA pixel coordinate; q. q.srIs prOne adjacent pixel coordinate of (a); epsilon represents a four-connected neighborhood set; w (p)r)＝(u(pr),v(pr) ) represents prSIFT-Flow displacement vector of (1); u and v represent the horizontal and vertical components of the SIFT-Flow displacement vector, respectively; d is a threshold value, with a default value of 40; α is a weighting factor for the second term, with a default value of 2;
the step S4 specifically includes the following steps:
step S41: calculating the aspect ratio similarity of the corresponding local blocks between the original image and the display adaptive image by using the pixel-level mapping relation, so as to calculate the geometric distortion of the local blocks, wherein the calculation formula is as follows:
Figure FDA0003623317920000021
wherein k represents the index sequence number of the local block in the original image; the ratio of variation of the width and height of the local block is rw(k)＝w*(k) N and rh(k)＝h*(k)/N；w*(k) And h*(k) Width and height, respectively, of local blocks of the reconstructed display-adapted image, obtained by calculating the maximum horizontal and vertical distances between pixels in the local blocks; when the variation ratio of width and height is the same, the local block equivalent to the original image is scaled in equal proportion, and the AR measure reaches the maximum value, i.e. SAR＝1；
Step S42: the area similarity of the corresponding local blocks is calculated by utilizing the pixel-level mapping relation, so that the information loss of the local blocks is measured, and the calculation formula is as follows:
Figure FDA0003623317920000022
wherein r isa(k)＝a*(k)/N2Is the area change ratio of the kth local block; eta is a positive number with a default value of 0.3, which is used to balance geometric distortion and information loss in the image fidelity measureA weight value;
step S43: and calculating the image fidelity of the local block displaying the adaptive image by using the aspect ratio similarity and the area similarity, wherein the calculation formula is as follows:
F(k)＝SAR(k)·SA(k)；
the step S5 specifically includes the following steps:
step S51: defining distortion visibility of each local block according to the average value of each local block in the saliency map S, and calculating the objective quality score of each block according to the fidelity and the distortion visibility of the block:
Figure FDA0003623317920000031
wherein the content of the first and second substances,
Figure FDA0003623317920000032
an average saliency value representing the kth local block of the saliency map S; if image fidelity F (k) is normalized to [0,1 ]]Then the mass range of the kth local block is
Figure FDA0003623317920000033
Therefore, when
Figure FDA0003623317920000034
The mass range of the corresponding local block is [0,1 ]](ii) a And following with
Figure FDA0003623317920000035
The mass base number of each local block is continuously increased, and the range is continuously reduced;
step S52: obtaining an objective quality score of the whole display adaptation image by calculating an average value of the quality scores of all the local blocks; calculating the objective quality score of each local block according to the fidelity and distortion visibility of the block:
Figure FDA0003623317920000036
wherein Q isNBPRepresenting the objective quality score calculated by the visible distortion pooling strategy, and M representing the number of local blocks in the original image.
2. The image display adaptation quality assessment method based on visible distortion pooling of claim 1, wherein: in step S51, the defining the distortion visibility of each local block according to the average value of each local block in the saliency map S specifically includes: defining a mass range with small cardinality and large interval for local blocks with high significance; meanwhile, a quality range with large base number and small interval is defined for the local blocks with low significance.
CN201910237280.0A 2019-03-27 2019-03-27 Image display adaptation quality evaluation method based on visible distortion pooling Active CN109978859B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910237280.0A CN109978859B (en) 2019-03-27 2019-03-27 Image display adaptation quality evaluation method based on visible distortion pooling

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910237280.0A CN109978859B (en) 2019-03-27 2019-03-27 Image display adaptation quality evaluation method based on visible distortion pooling

Publications (2)

Publication Number Publication Date
CN109978859A CN109978859A (en) 2019-07-05
CN109978859B true CN109978859B (en) 2022-06-14

Family

ID=67080897

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910237280.0A Active CN109978859B (en) 2019-03-27 2019-03-27 Image display adaptation quality evaluation method based on visible distortion pooling

Country Status (1)

Country Link
CN (1) CN109978859B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111127460B (en) * 2019-12-31 2022-06-14 福州大学 Image display adaptation evaluation method based on block connection
CN111242916B (en) * 2020-01-09 2022-06-14 福州大学 Image display adaptation evaluation method based on registration confidence measurement

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106780452A (en) * 2016-12-07 2017-05-31 华侨大学 A kind of combination small echo and the complete of Corner Feature refer to screen image quality measure method
CN106991669A (en) * 2017-03-14 2017-07-28 北京工业大学 A kind of conspicuousness detection method based on depth-selectiveness difference

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3042498A4 (en) * 2013-09-06 2017-11-08 Zhou Wang Method and system for objective perceptual video quality assessment

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106780452A (en) * 2016-12-07 2017-05-31 华侨大学 A kind of combination small echo and the complete of Corner Feature refer to screen image quality measure method
CN106991669A (en) * 2017-03-14 2017-07-28 北京工业大学 A kind of conspicuousness detection method based on depth-selectiveness difference

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
Backtracking Spatial Pyramid Pooling-Based Image Classifier for Weakly Supervised Top–Down Salient Object Detection;Hisham Cholakkal 等;《IEEE Transactions on Image Processing》;20180810;6064 - 6078 *
基于显著性图像边缘的全参考图像质量评价;闫钧华等;《仪器仪表学报》;20160915;第37卷(第09期);216-224 *
基于视觉特性和自然场景统计特性的图像质量评价研究;贾惠珍;《中国博士学位论文全文数据库 信息科技辑》;20180731;I138-59 *

Also Published As

Publication number Publication date
CN109978859A (en) 2019-07-05

Similar Documents

Publication Publication Date Title
CN107578404B (en) View-based access control model notable feature is extracted complete with reference to objective evaluation method for quality of stereo images
CN116205823B (en) Ultrasonic image denoising method based on spatial domain filtering
CN114972329A (en) Image enhancement method and system of surface defect detector based on image processing
CN101976444B (en) Pixel type based objective assessment method of image quality by utilizing structural similarity
CN110689000B (en) Vehicle license plate recognition method based on license plate sample generated in complex environment
CN111598869B (en) Method, equipment and storage medium for detecting Mura of display screen
CN102081742B (en) Method for automatically evaluating writing ability
CN109978859B (en) Image display adaptation quality evaluation method based on visible distortion pooling
CN103177451A (en) Three-dimensional matching algorithm between adaptive window and weight based on picture edge
CN109544504A (en) Screen picture quality evaluating method based on rarefaction representation
US20180253885A1 (en) Information processing apparatus and information processing method
CN112950596A (en) Tone mapping omnidirectional image quality evaluation method based on multi-region and multi-layer
CN104574312A (en) Method and device of calculating center of circle for target image
US20130083993A1 (en) Image processing device, image processing method, and program
CN101937562A (en) Construction method for gray-level information content histogram
CN115797607A (en) Image optimization processing method for enhancing VR real effect
CN112132774A (en) Quality evaluation method of tone mapping image
CN105631849B (en) The change detecting method and device of target polygon
CN109523590A (en) A kind of 3D rendering depth information visual comfort appraisal procedure based on sample
CN107578406A (en) Based on grid with Wei pool statistical property without with reference to stereo image quality evaluation method
US20230386023A1 (en) Method for detecting medical images, electronic device, and storage medium
CN114494262B (en) Method and device for evaluating image contrast
CN111369435A (en) Color image depth up-sampling method and system based on self-adaptive stable model
CN115601267A (en) Global tone mapping method with local detail compensation capability
CN111083468B (en) Short video quality evaluation method and system based on image gradient

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant